Statistical understanding for snow cover effects on near-surface ground temperature at the margin of maritime Antarctica, King George Island

Abstract

Snow cover plays an important role in water supply through melting of snow/ice in polar ecosystem and environments, in particular, Antarctica. Although a site access to Antarctica is high-priced, measurements of ground surface temperature (GST) using small self-recording temperature sensors (iButtons) can provide a powerful and relatively inexpensive approach to trace the spatial and temporal distributions of soil temperature and in addition, absence/presence of snow cover. In this study, principal component analysis (PCA) was utilized to explain major patterns of GST from 128 sites at King George Island in maritime Antarctica. Variations of GST were monitored between December 2011 and January 2013. The iButtons were initially installed in snow free areas in the austral summer of 2011. Principal components 1 and 2 were associated with air temperature and snow cover, respectively. Both PCs showed good correlations with the mean GST of JJA (June to August), not with that of DJF (December to January). Based on the results excluding an outlier, PCA divided the 127 GST observations into three groups effectively: absence of snow cover (Group 1), intermittent snow cover (Group 2), and presence of snow cover (Group 3). Snow cover can supply water to the ecosystem and the GST pattern of Group 2 may imply an abundance and high productivity of mosses in the study area. Using approaches suggested by previous studies, snow cover showed up nine days (days 317-326), and the melt-out date was day 326 (KG125 in Group 2). The GST data and statistical approaches used in this study can be useful in other GST studies, particularly, for both polar regions.